Prediction-Based Prefetch Scheduling in Mobile Service Applications

The usage of Web or Cloud-based applications on mobile devices suffers from varying link quality, which causes user-perceivable delays and therefore reduces the Quality of Experience (QoE). On the other hand, mobile devices are increasingly feature-rich, allowing to make usage of context data in order to predict network quality based on the user's location. In this paper, we propose a prefetch scheduling algorithm based on network quality predictions, and evaluate it using data collected from real-world field tests. We show that our approach fulfills the expected gain in QoE. Using network quality prediction to optimize data prefetching can improve the user-perceived response time by up to 95%.Our results not only show the feasibility of the proposed algorithm, but also motivate further research in the field of mobility pattern creation, and undermine the importance of location as a part of user context throughout the software stack.